Exploring Features For Localized Detection of Speech Recognition Errors

We address the problem of localized error detection in Automatic Speech Recognition (ASR) output to support the generation of tar-geted clarifications in spoken dialogue sys-tems. Localized error detection finds specific mis-recognized words in a user utterance. Tar-geted clarifications, in contrast with generic ‘please repeat/rephrase ’ clarifications, target a specific mis-recognized word in an utter-ance (Stoyanchev et al., 2012a) and require accurate detection of such words. We extend and modify work presented in (Stoyanchev et al., 2012b) by experimenting with a new set of features for predicting the likelihood of a local error in an ASR hypothesis on an un-sifted version of the original dataset. We im-prove over baseline results, where only ASR-generated features are used, by constructing optimal feature sets for utterance and word mis-recognition prediction. The f-measure for identifying incorrect utterances improves by 2.2 % and by 3.9 % for identifiying incorrect words.